The goal of our research is to study the signaling molecules operative in the initiation and progression of breast cancer. To this end, we have used human and mouse systems, genomic and cell biologic methods, and clinical as well as molecular approaches in interrogating these questions. Our clinically based population studies with CALGB determined that node positive breast cancers harboring aberrations of the HER-2 locus can achieve a doubling of survival if treated with dose-intensive doxorubicin regimens. This finding has been corroborated by the NSABP. We found a similar result with ras mutations and response to dose-intensive Ara-C in de novo AML: ras mutation positive AML exhibit a steeper dose-response to Ara-C than ras mutation negative AMLs. These results suggested that specific oncogene abnormalities can render tumor cells responsive to certain therapeutic interventions. Given that signaling molecules can modulate the clinical behavior of important tumors, we have investigated the biologiy of specific signalling molecules potentially operative in human breast cancers. In the realm of known signalling molecules, we have determined that HER-2 amplification is found in the earliest forms of breast cancer such as DCIS, and in benign breast disease as well. Though the frequency of amplification is low in benign histologies, its presence along with hyperplasia raises the risk of subsequent cancer by 7 fold. Thus, HER-2 amplification is seen early in the genesis of breast cancers. With BRCA1, we have discovered that the putative tumor suppressor gene can inhibit the growth of cancer cells only in the presence of an intact pRB protein. In cell lines harboring mutant rb, no growth arrest is seen when BRCA1 is overexpressed. We have found that BRCA1 binds to hypophosphorylated pRB, through a 90 amino acid domain in the acidic portion of the protein. This raises the possibility that BRCA1 affects cell growth through the Rb pathway and that modulation of RB levels may change the biological effects of BRCA1. In the realm of previously unknown signalling molecules, we had identified a number of new kinases expressed in primary breast cancers. This work led to the discovery of JAK3 as the signaling partner of the IL-2 receptor and the overexpression of kinase deficient isoforms in breast cancer, to the characterization of the human CDK7, and the cloning of a novel NIMA-related serine-threonine kinase, STK2. More recently we have concentrated on two kinases because of their biological effects on epithelial cancer growth: AXL and RAK. AXL, a new class of receptor tyrosine kinase (RTK), is characterized by an extracellular domain resembling adhesion molecules. We have determined that Axl is rendered transforming by overexpression resulting in enforced oligomerization and activation of the ras/MAPK pathway. Axl is overexpressed in 30 percent of breast cancers and appears to be coexpressed with cathepsin D. Since proliferation and transformation were associated with conditional activation of the kinase function, we isolated and cloned the ligand for AXL and found it to be the growth arrest specific gene 6 or GAS-6, which is a homologue of the coagulation cofactor, Protein S. Our current work has found GAS-6 to be a motogenic factor especially in vascular smooth muscle cells, an survival factor, and a mediator of Axl-dependent cell adhesion, suggesting a role in the invasion and metastatic process. The motogenic function does not require ras activation but does require PI-3' kinase and AKT engagement. Thus, we are defining the minimal signaling cassette necessary for motogenesis. Interestingly, we have found that motogenesis and survival appear to be linked both biologically and biochemically. RAK is a new nuclear tyrosine kinase structurally related to the src/fyn family of kinases and expressed predominantly in epithelial tissues and cancers. We have determined that RAK is a potent growth inhibitory kinase localized to 6q in a region commonly deleted in breast and ovarian cancers. Uniquely, rak is a centrosomal protein that also resides in the Golgi when overexpressed suggesting that rak is involved in the centrosome-Golgi tether. Current work seeks to investigate the mechanism of growth inhibition through gene disruption studies and the identification of associated proteins. Both have been accomplished and analysis is underway. Given our interest in new technologies for interrogating cell biology, we have developed the capacity to manufacture and to analyze cDNA microarrays. To date, we have tested 6,000 cDNA human, 2,200 cDNA mouse, and 4,000 cDNA rat arrays with success. Our results show that primary tumors can be classified according to the array profiles, that null phenotypes of disrupted oncogenes can exhibit significant expression changes as detected by microarrays, and that specific pathways activate and suppress the same common genes. We have explored the expression "footprints" of cells after exposure to selected therapeutic agents, and have observed stereotyped expression profiles in normal mammary gland development. Specifically, we have determined the myc induces the expression of protein synthesis genes as its primary role, and have identified the engagement of the wnt pathway in thyroid hormone function. We have developed approaches to derive expression profiles using RNA from needle aspirates, thus opening the possibility of tracking temporal changes in gene expression from clinical samples. This then provides the linkage between the cell biology in the laboratory and the clinical work by our section through studying the effects of therapeutic interventions on gene expression patterns.